So what is correct here?
Well I do agree that "AI safety" in media and large company discourse has been heavily influenced by Yudkowsky's ideas and that those ideas focus on stuff like the "paperclip maximizer" issue. This is the crux of the problem.
bsky.app/profile/segy...
@segyges.bsky.socialYudkowsky's most viral idea, the one that infects the most people and spreads the farthest, is that AI, if we continue to develop it, will inevitably 1) achieve superintelligence ("RSI"), 2) escape human control 2) kill everyone ("paperclip them", "doom")